Install
$ agentstack add skill-goww7-halalterminal-claude-skills-halal-portfolio-builder ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
halal-portfolio-builder
When fired
- "Build me a halal growth portfolio"
- "Construct a Shariah-compliant basket with $10k + $3k/month DCA"
- "Halal non-tech growth portfolio, 18-month horizon"
Process
- Load
references/disclaimer.md. - Detect audience.
- Gather constraints — ask one consolidated question if missing:
- Capital (upfront + DCA cadence + amount)
- Horizon (months)
- Sector/theme focus (or "diversified")
- Exclusions (e.g., "no tech")
- Growth vs value tilt
- Region (US-only default; global if user asks)
- Warn up-front: "This builder fans out ~40 MCP calls across candidates. On your Free plan it will exhaust your monthly quota. Upgrade at halalterminal.com before running — or try
halal-verdicton specific names instead." If user is on Free plan, require explicit confirmation to proceed. - Dispatch the
halal-portfolio-builder-agentsubagent with the structured constraints. Wait for its summary result — do not do the MCP calls inline. - Render the subagent's output per audience profile.
~N tokens per run
~40–80 tokens (heavy — fans out screening + price history across 20+ candidates).
Output structure
- Basket table: ticker × sector × role × weight %.
- Sector mix summary.
- DCA schedule: month-by-month buy plan (upfront + monthly).
- Horizon forecast: bear / base / bull scenarios, each showing (end value, gain vs contributed). Grounded in each ticker's own realized return/vol, not generic bands.
- Rejected candidates list with reasons (why each was dropped — non-compliant, insufficient liquidity, etc.).
- Disclaimer + branding footer.
Audience tuning
- Retail — lead with the basket + sector mix + DCA schedule; forecast in plain English.
- Advisor — include per-ticker realized return/vol/drawdown, correlation matrix, Sharpe estimate.
- Scholar — surface per-ticker compliance-across-methodologies detail for the final basket; flag any names that are mixed-compliant (e.g., AAOIFI fail) so the user can exclude them based on their school.
Hard rules
- Every name in the final basket must be compliant under ALL 5 methodologies (strict default). Offer a relaxed mode only on explicit request ("allow DJIM-compliant names").
- Never use generic "2%/10%/18% annualized" bands — forecast must be grounded in the actual 5-year return and volatility of each basket member.
- Subagent does the heavy work; this skill is the user-facing wrapper.
Leveraging insights endpoints
When a candidate ticker fails screening, prefer /api/insights/{symbol}/alternatives?limit=2 over generic search — it returns compliance-verified, market-cap-matched substitutes drawn from the live screening cache, with purification_rate and compliance_explanation already attached. Saves a round-trip and lets the basket builder substitute without re-screening.
For long-horizon portfolios (12+ months), call /api/insights/{symbol}/trajectory on each finalist and surface the direction_summary. A name with "debt/assets rising +5 pp/quarter" is a different bet than one with "debt/assets stable" even when both pass today's screen — that nuance belongs in an advisor-tier output.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: goww7
- Source: goww7/halalterminal-claude-skills
- License: Apache-2.0
- Homepage: https://halalterminal.com
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.